Hugging Face Compumacy/Psych_Qwen_32B: Specialized Clinical Psychology LLM
Compumacy/Psych_Qwen_32B, developed by Daemontatox, is an experimental 32 billion parameter large language model built on unsloth/Qwen3-32B. It is specifically fine-tuned for applications in clinical psychology and psychiatry, aiming to assist mental health professionals with structured assessments and treatment planning.
Key Capabilities
This model excels at emulating the reasoning process of a psychiatrist, providing structured clinical documentation. Its core capabilities include:
- Chief Complaint Analysis: Objectively summarizing presenting symptoms from clinical vignettes.
- Differential Diagnosis: Systematically evaluating potential diagnoses against DSM-5-TR criteria and referencing ICD-11 classifications.
- Comprehensive Risk Assessment: Identifying and assessing risks such as suicide, homicide, psychosis, and substance use.
- Evidence-Based Recommendations: Proposing treatment plans (pharmacotherapy, psychotherapy) aligned with guidelines from organizations like APA, WFSBP, and NICE.
- Monitoring and Referrals: Outlining necessary follow-ups and indications for specialized care.
Training and Use
The model was fine-tuned using the Unsloth framework on the Daemontatox/Psy-Data-books dataset, which comprises professional psychiatric literature and clinical guidelines. It requires a specific Alpaca-style prompt structure for optimal performance and is recommended for use with the Unsloth library for efficient inference. It is crucial to note that this model is a research tool intended to augment, not replace, the expertise of licensed clinicians, and all outputs require independent verification due to potential biases and the risk of hallucination.